parameter.search.parallel: Parameter Search for Parallel Processing

In tbrycekelly/TheSource: TheSource: It Will Hold Your Beer

Parameter Search for Parallel Processing

Description

Implements a recursive grid search routine to solve optimization problems in arbitrary dimensions.

Usage

parameter.search.parallel(
  n,
  cost,
  grid = NULL,
  ...,
  bounds,
  splits = 10,
  progression = NULL
)

Arguments

n	Number of recursions to perform
cost	The cost function which must return a numeric value and accept parameter values as the first arguments and in the order they are provided.
...	Optional argument that is passed directly onto the cost function
bounds	A dataframe containing the minimum and maximum values permitted of each parameter
splits	The number of subdivisions to perform for each dimension (so grid size is n x splits ^ dimensionality)
progression	The size of the new search-space to interrogate. A value between 1 and splits/2. Default value (NULL) will yeield a progression of max(1, splits/4), good for most problems.

Author(s)

Thomas Bryce Kelly

tbrycekelly/TheSource documentation built on Nov. 7, 2023, 12:48 a.m.